Bayesian Deep Learning (opens in new tab)

(twiecki.github.io)

7 pointstwiecki9y ago3 comments

3 comments

Very nice article!

But ... You generated your original data from: "sklearn.datasets.make_moons". Then you say that you tested your classifier on a hold-out set with "sample_ppc()".

I don't think that's a true hold-out set. A true hold-out set would be obtained by running "make_moons()" again to generate new data.

Have you tried that? It would be interesting I think.

I'll run the test myself too ...

Again, thanks for a very nice and informative article.

twieckiOP9y ago

Thanks for your comment! I did split the data in two, but it's easy to miss. X_test and X_train are the two sets. ann_input.set_value(X_test) then switches in the test values. That's identical to running make_moons() again.

lotsoflumens9y ago

OK, I didn't understand how the "neural_network" was connected to the data.

Thanks for responding.

j / k navigate · click thread line to collapse

3 comments

lotsoflumens9y ago

Very nice article!

But ... You generated your original data from: "sklearn.datasets.make_moons". Then you say that you tested your classifier on a hold-out set with "sample_ppc()".

I don't think that's a true hold-out set. A true hold-out set would be obtained by running "make_moons()" again to generate new data.

Have you tried that? It would be interesting I think.

I'll run the test myself too ...

Again, thanks for a very nice and informative article.

twieckiOP9y ago

lotsoflumens9y ago

OK, I didn't understand how the "neural_network" was connected to the data.

Thanks for responding.

j / k navigate · click thread line to collapse